Dataset Information

Bioacoustic classification of avian calls from raw sound waveforms with an open-source deep learning architecture.

ABSTRACT: The use of autonomous recordings of animal sounds to detect species is a popular conservation tool, constantly improving in fidelity as audio hardware and software evolves. Current classification algorithms utilise sound features extracted from the recording rather than the sound itself, with varying degrees of success. Neural networks that learn directly from the raw sound waveforms have been implemented in human speech recognition but the requirements of detailed labelled data have limited their use in bioacoustics. Here we test SincNet, an efficient neural network architecture that learns from the raw waveform using sinc-based filters. Results using an off-the-shelf implementation of SincNet on a publicly available bird sound dataset (NIPS4Bplus) show that the neural network rapidly converged reaching accuracies of over 65% with limited data. Their performance is comparable with traditional methods after hyperparameter tuning but they are more efficient. Learning directly from the raw waveform allows the algorithm to select automatically those elements of the sound that are best suited for the task, bypassing the onerous task of selecting feature extraction techniques and reducing possible biases. We use publicly released code and datasets to encourage others to replicate our results and to apply SincNet to their own datasets; and we review possible enhancements in the hope that algorithms that learn from the raw waveform will become useful bioacoustic tools.

SUBMITTER: Bravo Sanchez FJ

PROVIDER: S-EPMC8333097 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Bioacoustic sensors, sometimes known as autonomous recording units (ARUs), can record sounds of wildlife over long periods of time in scalable and minimally invasive ways. Deriving per-species abundance estimates from these sensors requires detection, classification, and quantification of animal vocalizations as individual acoustic events. Yet, variability in ambient noise, both over time and across sensors, hinders the reliability of current automated systems for sound event detection (SED), such as convolutional neural networks (CNN) in the time-frequency domain. In this article, we develop, benchmark, and combine several machine listening techniques to improve the generalizability of SED models across heterogeneous acoustic environments. As a case study, we consider the problem of detecting avian flight calls from a ten-hour recording of nocturnal bird migration, recorded by a network of six ARUs in the presence of heterogeneous background noise. Starting from a CNN yielding state-of-the-art accuracy on this task, we introduce two noise adaptation techniques, respectively integrating short-term (60 ms) and long-term (30 min) context. First, we apply per-channel energy normalization (PCEN) in the time-frequency domain, which applies short-term automatic gain control to every subband in the mel-frequency spectrogram. Secondly, we replace the last dense layer in the network by a context-adaptive neural network (CA-NN) layer, i.e. an affine layer whose weights are dynamically adapted at prediction time by an auxiliary network taking long-term summary statistics of spectrotemporal features as input. We show that PCEN reduces temporal overfitting across dawn vs. dusk audio clips whereas context adaptation on PCEN-based summary statistics reduces spatial overfitting across sensor locations. Moreover, combining them yields state-of-the-art results that are unmatched by artificial data augmentation alone. We release a pre-trained version of our best performing system under the name of BirdVoxDetect, a ready-to-use detector of avian flight calls in field recordings.

Project description:BackgroundBeing able to accurately assess the 3D architecture of plant canopies can allow us to better estimate plant productivity and improve our understanding of underlying plant processes. This is especially true if we can monitor these traits across plant development. Photogrammetry techniques, such as structure from motion, have been shown to provide accurate 3D reconstructions of monocot crop species such as wheat and rice, yet there has been little success reconstructing crop species with smaller leaves and more complex branching architectures, such as chickpea.ResultsIn this work, we developed a low-cost 3D scanner and used an open-source data processing pipeline to assess the 3D structure of individual chickpea plants. The imaging system we developed consists of a user programmable turntable and three cameras that automatically captures 120 images of each plant and offloads these to a computer for processing. The capture process takes 5-10 min for each plant and the majority of the reconstruction process on a Windows PC is automated. Plant height and total plant surface area were validated against "ground truth" measurements, producing R2 > 0.99 and a mean absolute percentage error < 10%. We demonstrate the ability to assess several important architectural traits, including canopy volume and projected area, and estimate relative growth rate in commercial chickpea cultivars and lines from local and international breeding collections. Detailed analysis of individual reconstructions also allowed us to investigate partitioning of plant surface area, and by proxy plant biomass.ConclusionsOur results show that it is possible to use low-cost photogrammetry techniques to accurately reconstruct individual chickpea plants, a crop with a complex architecture consisting of many small leaves and a highly branching structure. We hope that our use of open-source software and low-cost hardware will encourage others to use this promising technique for more architecturally complex species.

Dataset Information

Bioacoustic classification of avian calls from raw sound waveforms with an open-source deep learning architecture.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets